Keep a history of selected Algorithms and datasets around in the toolchain
With the new editor, it's only possible to add blocks based on available datasets, algorithms, and analyzers. Which might be a good thing according to #265 (closed). Also, once blocks are added (according to the algorithm or dataset it was selected from), it is no longer associated with the block that it was chosen from. This makes sense because as far as the toolchain is concerned, this association is irrelevant and this association is kept around in the experiment definition.
However, I think it makes sense to keep association around in the toolchain. Doing so can let us:
- Provide hints to the user (reminding them what algorithm that block was) so that they have to memorize less when designing the toolchain.
- Automatically figure out the channel of blocks as disccussed in #272.
On the other hand, we want to merge the toolchain/experiment definitions per https://gitlab.idiap.ch/beat/beat.core/-/issues/88
So maybe this issue must be looked at when that works is already done?!